Hedging in energy markets

Dublin, 6 March 2018

www.kyos.com, +31 (0)23 5510221Cyriel de Jong, dejong@kyos.com

• Comparing NL-DE with IE-GB markets

• Some general hedging principles

• Proxy hedging

• Minimizing VaR versus EaR

• Hedge performance in NL market with DE forwards

• Conclusion

Outline

2

• Especially 1 large neighbouring country: Germany (vs GB)

• Interconnection capacitieshave been growing

• NL-DE: from about2 to 4 GW

• NL-DE: market coupling sinceNov 2010

• NL: also connections toBE, GB, NO

Comparison Ireland with the Netherlands

3

• The Netherlands has been a net importer of German power

• Nuclear phase-out in Germany and coal/lignite plant closureswill lead to a more balanced situation

• At the same time, transmission capacities will grow further

Interconnection flows expectation

4

Source: KYOS

fundamental

modelanalysis

• Variations in market structure

• E.g.: Energiewende Germany: massive growth in renewablesfrom 2011 onwards

• DE and NL markets ‘coupled’ since Nov-2010: most equalprices in 2011

Price differentials NL - DE

5

• German power market is the most liquid in Europe

• Dutch power market players can use German power forwards as a proxy hedge for their exposures

• What is a proxy hedge?

• Actual exposure is to prices in market A (= NL, or IE)

• Hedging is with instruments in market B (= DE, or GB)

• Proxy hedging effectiveness depends on:

• Correlations in price returns (short-term)

• Similarity in price movements (long-term)

DE power forwards good hedge for NL power?

6

Hedge result: risk reduction

7

Hedging

8

Hedge

A hedge is a position created to offset an exposure to price fluctuations in some opposite position (or market) with the goal of minimizing the exposure to unwanted risk

Power plant hedge

• Sell Power on forward market

• Buy fuels (gas/coal) on forward market

• Buy CO2 credits on forward market

• Don’t forget fx hedges (e.g. CO2 is traded in EUR, Power/Gas in GBP)

Remember: hedge is done to minimize market risk, not operational (technical) risk

1. When to hedge? Over what horizon?

• Today, next week, next month

• Everything at once, or gradually over time

2. Adjust hedge over time?

• Static or dynamic hedge

• Rolling intrinsic or delta

3. What products to use for hedging?

• Baseload, peakload

• Calendar, season, quarter, month

4. Volume or value?

Plant hedging: decisions

9

Optimal hedge ratio

Intuition:

High correlation (ρPF) between returns in your open position (P) and your hedge instrument (F), means a better hedge (it’s worth hedging)

High volatility of open position (σP) relative to volatility of hedge instrument (σF), means I need a sizeable hedge

Definition: h* = ρPF * σP / σF

Implication: if you hedge an exposure in market P (=NL, IE) with products of market B (=DE, GB) you should reduce the (value-based) hedge volume in proportion to the correlation. Correlation is not 100%!

10

• The hedge which minimizes the Value-at-Risk (VaR) is based on:

• value/delta hedge, and

• optimal hedge ratio

• Note: Value-at-Risk is the ‘maximum’ loss in market value over a short horizon (e.g. 1 day) with a certain confidence (e.g. 95%)

Optimal hedging to minimize VaR

11

Using forward price data from 2005 to2017:

• Volatility term structure similar

• Correlations lower close to maturity

• Forward correlation has increased

Analysis of NL and DE forward price returns

12

• VaR is a suitable risk measure in liquid markets, where market value can be accurately measured and managed

• VaR is less suitable for an asset portfolio, especially if there is significant spot price risk

• For example, if a power producer sells the expected power production (or delta exposure of the plant) forward:

• The ‘hope’ is that the expected plant value (today) is eventually realized (in the future)

• If revenues in the spot market are lower than expected by X million Euro, then this should be compensated by a forward hedging profit of X million Euro.

Minimizing VaR or other risk metric?

13

• EaR is similar to VaR, but:

• Focusses on realized earnings during the delivery period

• Minimizing EaR and VaR may lead to somewhat different hedgevolumes.

Earnings-at-Risk (EaR)

14

• We are 1 MWh long in year Y, in NL

• We sell x MWh forward in DE

• In year Y-1 we gradually sell the Y+1 forward product

• We assess P&L for different levels of x (= hedge ratio)

Historical analysis DE hedge for NL market

15

Optimal hedge ratio, historically: 50%

May be different over other horizons, and also when consideringother hedging strategies (e.g. involving quarterly and monthlyproducts, peak versus base, etc.)

Analysis continued

16

• Forward trading in IE market is likely to remain illiquid

• Proxy hedging with other products may be needed

• With fuels? Don’t trust (static) fundamental models!

• With power forwards in GB? Don’t overhedge!

• Try to assess optimal hedge volumes, and monitor performance over time

Conclusion

17